1 Preamble

1.1 R for data science

R for data science: The best place to start learning the tidyverse by Hadley Wickham and Garrett Grolemund (available online : https://r4ds.had.co.nz). ggplot2: elegant graphics for data science by Hadley Wickham. Goes into greater depth into the ggplot2 visualisation system.

2 Tidyverse & tidy data

2.1 Tidyverse

R packages for data science:

The tidyverse is an opinionated collection of R packages designed for data science. All packages share an underlying design philosophy, grammar, and data structures. (https://www.tidyverse.org/)

install.packages("tidyverse")

2.2 Tidyverse

require(tidyverse)
## Loading required package: tidyverse
## ── Attaching packages ─────────────────────────────────────────────── tidyverse 1.2.1 ──
## ✔ ggplot2 3.1.0     ✔ purrr   0.2.5
## ✔ tibble  2.0.1     ✔ dplyr   0.7.8
## ✔ tidyr   0.8.2     ✔ stringr 1.4.0
## ✔ readr   1.3.1     ✔ forcats 0.3.0
## ── Conflicts ────────────────────────────────────────────────── tidyverse_conflicts() ──
## ✖ dplyr::filter() masks stats::filter()
## ✖ dplyr::lag()    masks stats::lag()

2.3 Tidy data

_ Tidying: structuring datasets to facilitate analysis._

A tidy dataset :

2.4 Why tidying a dataset ?

Typical presentation dataset
treatment_a treatment_b treatment_c
John Smith 16 17 20
Jane Doe 23 25 23
Mary Johnson 15 15 19 
boxplot(treatments)

2.5 Why tidying a dataset ?

treatments.2 <- t(treatments)
The same data but structured differently
John Smith Jane Doe Mary Johnson
treatment_a 16 23 15
treatment_b 17 25 15
treatment_c 20 23 19 
boxplot(treatments.2)

2.6 Why tidying a dataset ?

An exemple of a tidy dataset
person treatment result
John Smith treatment_a 16
Jane Doe treatment_a 23
Mary Johnson treatment_a 15
John Smith treatment_b 17
Jane Doe treatment_b 25
Mary Johnson treatment_b 15
John Smith treatment_c 20
Jane Doe treatment_c 23
Mary Johnson treatment_c 19 
boxplot(result~treatment,data=treatments.3)

2.7 Why tidying a dataset ?

boxplot(result~person,data=treatments.3)

boxplot(result~person+treatment,data=treatments.3)

3 Pipe and tibble

3.1 Pipes

The pipe %>%:

  • Come from the package by Stefan Milton Bache.
  • Automatically loaded in tidyverse.
  • Equivalent to | in bash
cat iris.tsv | cut -f5 | sed 's/^./\U&/' | head
## Species
## Setosa
## Setosa
## Setosa
## Setosa
## Setosa
## Setosa
## Setosa
## Setosa
## Setosa
read_tsv("iris.tsv",col_names = F) %>% pull(5) %>% str_to_title() %>% head
## [1] "Species" "Setosa"  "Setosa"  "Setosa"  "Setosa"  "Setosa"

3.2 Pipes vs no pipes : without pipe

ex.dat <- rnorm(n = 1000,mean = 5,sd=1)
sub.dat <- sample(ex.dat,size = 100,replace=F)
hist(sub.dat)

3.3 Pipes vs no pipes : without pipe

hist(sample(rnorm(n = 1000,mean = 5,sd=1),size = 100,replace=F))

3.4 Pipes vs no pipes : with pipe

rnorm(n = 1000,mean = 5,sd=1) %>%
    sample(size = 100,replace=F) %>%
    hist()

3.5 Tibbles: tibble::tibble

my.tibble <- tibble(
    person = c("John Smith","Jane Doe","Mary Johnson"),
    treatment_a = sample(10:25,size = 3,replace = T),
    treatment_b = sample(10:25,size = 3,replace = T),
    treatment_c = sample(10:25,size = 3,replace = T)
)
## # A tibble: 3 x 4
##   person       treatment_a treatment_b treatment_c
##   <chr>              <int>       <int>       <int>
## 1 John Smith            18          12          15
## 2 Jane Doe              17          15          17
## 3 Mary Johnson          14          14          15

3.6 Tibbles: tibble::tibble()

Pros:

  • Cells can contain list and data.frame.
  • It never changes an input’s type (i.e., no more stringsAsFactors = FALSE!).
  • Can use not valid R variable names (:)) as column names.
  • Refined print method that shows only the first 10 rows.

Cons:

  • It never uses row.names().
    • use tibble::rownames_to_column().
treatments %>% tibble::rownames_to_column("person") %>%
    as_tibble() %>%
    rename(`:)` = person)
## # A tibble: 3 x 4
##   `:)`         treatment_a treatment_b treatment_c
##   <chr>              <int>       <int>       <int>
## 1 John Smith            16          17          20
## 2 Jane Doe              23          25          23
## 3 Mary Johnson          15          15          19

3.7 Split a dataset into multiples dataset using tidyr::nest()

my.tibble.split <- my.tibble %>% group_by(person) %>% nest(.key = Data)
my.tibble.split
## # A tibble: 3 x 2
##   person       Data            
##   <chr>        <list>          
## 1 John Smith   <tibble [1 × 3]>
## 2 Jane Doe     <tibble [1 × 3]>
## 3 Mary Johnson <tibble [1 × 3]>
my.tibble.split[["Data"]][[1]]
## # A tibble: 1 x 3
##   treatment_a treatment_b treatment_c
##         <int>       <int>       <int>
## 1          18          12          15

3.8 Read/write tibbles: readr::read_* / readr::write_*

Read/write tibbles: :

  • readr::read_*:

    • read_csv: comma delimited files.
    • read_csv2: semi-colon delimited files.
    • read_tsv: tab delimited files.
    • read_delim: any delimiter.

  • readr::write_*:

    • write_csv: comma delimited files.
    • write_csv2: semi-colon delimited files.
    • write_tsv: tab delimited files.
    • write_delim: any delimiter.

3.9 Read/write tibbles: readr::read_* / readr::write_*

read_tsv("iris.tsv",
         col_types = cols(
            Sepal.Length = col_double(),Sepal.Width = col_double(),
            Petal.Length = col_double(),Petal.Width = col_double(),
            Species = col_character()
         )
    )
## # A tibble: 150 x 5
##    Sepal.Length Sepal.Width Petal.Length Petal.Width Species
##           <dbl>       <dbl>        <dbl>       <dbl> <chr>  
##  1          5.1         3.5          1.4         0.2 setosa 
##  2          4.9         3            1.4         0.2 setosa 
##  3          4.7         3.2          1.3         0.2 setosa 
##  4          4.6         3.1          1.5         0.2 setosa 
##  5          5           3.6          1.4         0.2 setosa 
##  6          5.4         3.9          1.7         0.4 setosa 
##  7          4.6         3.4          1.4         0.3 setosa 
##  8          5           3.4          1.5         0.2 setosa 
##  9          4.4         2.9          1.4         0.2 setosa 
## 10          4.9         3.1          1.5         0.1 setosa 
## # … with 140 more rows

4 Tidying dataset

4.1 Using tidyr to reshape data

Use tidyr::gather() and tidyr::spread() to reorganize the value.

treatments %>% tibble::rownames_to_column("person") %>%
    tidyr::gather(key="treatment",value = "result",-person)
##         person   treatment result
## 1   John Smith treatment_a     16
## 2     Jane Doe treatment_a     23
## 3 Mary Johnson treatment_a     15
## 4   John Smith treatment_b     17
## 5     Jane Doe treatment_b     25
## 6 Mary Johnson treatment_b     15
## 7   John Smith treatment_c     20
## 8     Jane Doe treatment_c     23
## 9 Mary Johnson treatment_c     19
treatment.3 %>% spread(key = "treatment",value = "result")
##         person treatment_a treatment_b treatment_c
## 1     Jane Doe          23          25          23
## 2   John Smith          16          17          20
## 3 Mary Johnson          15          15          19

4.2 Using tidyr to reshape data

treatment.3 %>% separate(person,into = c("First Name","Last Name"),sep=" ")
##   First Name Last Name   treatment result
## 1       John     Smith treatment_a     16
## 2       Jane       Doe treatment_a     23
## 3       Mary   Johnson treatment_a     15
## 4       John     Smith treatment_b     17
## 5       Jane       Doe treatment_b     25
## 6       Mary   Johnson treatment_b     15
## 7       John     Smith treatment_c     20
## 8       Jane       Doe treatment_c     23
## 9       Mary   Johnson treatment_c     19

5 Manipulate a dataset using dplyr

5.1 dplyr

dplyr:

  • Manipulate variables.
  • Manipulate observations.
  • Group / summarise observations.
iris dataset
Sepal.Length Sepal.Width Petal.Length Petal.Width Species
5.1 3.5 1.4 0.2 setosa
4.9 3.0 1.4 0.2 setosa
4.7 3.2 1.3 0.2 setosa
4.6 3.1 1.5 0.2 setosa
5.0 3.6 1.4 0.2 setosa
5.4 3.9 1.7 0.4 setosa

5.1.1 iris dataset

The famous (Fisher’s or Anderson’s) iris data set gives the measurements in centimeters of the variables sepal length and width and petal length and width, respectively, for 50 flowers from each of 3 species of iris. The species are Iris setosa, versicolor, and virginica. (from ?iris)

5.2 Manipulate variables using dplyr::select()

iris.tbl %>% select(1:3)
iris.tbl %>% select(Sepal.Length,Sepal.Width,Petal.Length)
iris.tbl %>% select(Sepal.Length:Petal.Length)
iris.tbl %>% select(-Petal.Width,-Species)
## # A tibble: 150 x 3
##    Sepal.Length Sepal.Width Petal.Length
##           <dbl>       <dbl>        <dbl>
##  1          5.1         3.5          1.4
##  2          4.9         3            1.4
##  3          4.7         3.2          1.3
##  4          4.6         3.1          1.5
##  5          5           3.6          1.4
##  6          5.4         3.9          1.7
##  7          4.6         3.4          1.4
##  8          5           3.4          1.5
##  9          4.4         2.9          1.4
## 10          4.9         3.1          1.5
## # … with 140 more rows

5.3 Manipulate variables using dplyr::select()

iris.tbl %>% select(starts_with("Petal"))
## # A tibble: 150 x 2
##    Petal.Length Petal.Width
##           <dbl>       <dbl>
##  1          1.4         0.2
##  2          1.4         0.2
##  3          1.3         0.2
##  4          1.5         0.2
##  5          1.4         0.2
##  6          1.7         0.4
##  7          1.4         0.3
##  8          1.5         0.2
##  9          1.4         0.2
## 10          1.5         0.1
## # … with 140 more rows

5.4 Manipulate variables using dplyr::pull()

iris.tbl %>% select(Sepal.Length)
## # A tibble: 150 x 1
##   Sepal.Length
##          <dbl>
## 1          5.1
## 2          4.9
## 3          4.7
## 4          4.6
## 5          5  
## # … with 145 more rows
iris.tbl %>% pull(Sepal.Length)
## [1] 5.1 4.9 4.7 4.6 5.0

5.5 Modify variables using dplyr::mutate()

iris.tbl %>%
    mutate(Petal.Length = 
               scales::percent(Petal.Length/max(Petal.Length))
           )
## # A tibble: 150 x 5
##    Sepal.Length Sepal.Width Petal.Length Petal.Width Species
##           <dbl>       <dbl> <chr>              <dbl> <fct>  
##  1          5.1         3.5 20.3%                0.2 setosa 
##  2          4.9         3   20.3%                0.2 setosa 
##  3          4.7         3.2 18.8%                0.2 setosa 
##  4          4.6         3.1 21.7%                0.2 setosa 
##  5          5           3.6 20.3%                0.2 setosa 
##  6          5.4         3.9 24.6%                0.4 setosa 
##  7          4.6         3.4 20.3%                0.3 setosa 
##  8          5           3.4 21.7%                0.2 setosa 
##  9          4.4         2.9 20.3%                0.2 setosa 
## 10          4.9         3.1 21.7%                0.1 setosa 
## # … with 140 more rows

5.6 Create variables using dplyr::mutate()

iris.tbl %>%
    #Case 1
    mutate(Size = dplyr::case_when(
            Petal.Length >= 4 & Petal.Width >= 1.3 ~ "Big",
            Petal.Length < 4 & Petal.Width < 1.3 ~ "Small"
        )
    ) %>%
    #Case 2
    mutate(Ratio_Petal_Sepal_width = Petal.Width/Sepal.Width)
## # A tibble: 3 x 7
##   Ratio_Petal_Sep… Size  Sepal.Length Sepal.Width Petal.Length Petal.Width
##              <dbl> <chr>        <dbl>       <dbl>        <dbl>       <dbl>
## 1           0.0571 Small          5.1         3.5          1.4         0.2
## 2           0.0667 Small          4.9         3            1.4         0.2
## 3           0.0625 Small          4.7         3.2          1.3         0.2
## # … with 1 more variable: Species <fct>

other functions:

  • dplyr::mutate_all(): Apply a function to every columns.
  • dplyr::mutate_at(): Apply a function to specific column.

5.7 Arrange dataset using dplyr::arrange()

iris.tbl %>% arrange(Petal.Length) %>% slice(1:5)
## # A tibble: 5 x 5
##   Sepal.Length Sepal.Width Petal.Length Petal.Width Species
##          <dbl>       <dbl>        <dbl>       <dbl> <fct>  
## 1          4.6         3.6          1           0.2 setosa 
## 2          4.3         3            1.1         0.1 setosa 
## 3          5.8         4            1.2         0.2 setosa 
## 4          5           3.2          1.2         0.2 setosa 
## 5          4.7         3.2          1.3         0.2 setosa
iris.tbl %>% filter(min_rank(Petal.Length) <= 5)
iris.tbl %>% top_n(-5,Petal.Length)

Use desc() to order by high to low.

5.8 Manipulate observations using dplyr::filter()

iris.tbl %>% filter(Sepal.Length < 7) %>% slice(1:5)
## # A tibble: 5 x 5
##   Sepal.Length Sepal.Width Petal.Length Petal.Width Species
##          <dbl>       <dbl>        <dbl>       <dbl> <fct>  
## 1          5.1         3.5          1.4         0.2 setosa 
## 2          4.9         3            1.4         0.2 setosa 
## 3          4.7         3.2          1.3         0.2 setosa 
## 4          4.6         3.1          1.5         0.2 setosa 
## 5          5           3.6          1.4         0.2 setosa
iris.tbl %>% filter(Species ==  "setosa") %>% slice(1:5)
## # A tibble: 5 x 5
##   Sepal.Length Sepal.Width Petal.Length Petal.Width Species
##          <dbl>       <dbl>        <dbl>       <dbl> <fct>  
## 1          5.1         3.5          1.4         0.2 setosa 
## 2          4.9         3            1.4         0.2 setosa 
## 3          4.7         3.2          1.3         0.2 setosa 
## 4          4.6         3.1          1.5         0.2 setosa 
## 5          5           3.6          1.4         0.2 setosa

5.9 Manipulate observations using dplyr::filter()

iris.tbl %>%
    filter(Sepal.Length < 7) %>%
    filter(Species ==  "setosa") %>% slice(1:5)
## # A tibble: 5 x 5
##   Sepal.Length Sepal.Width Petal.Length Petal.Width Species
##          <dbl>       <dbl>        <dbl>       <dbl> <fct>  
## 1          5.1         3.5          1.4         0.2 setosa 
## 2          4.9         3            1.4         0.2 setosa 
## 3          4.7         3.2          1.3         0.2 setosa 
## 4          4.6         3.1          1.5         0.2 setosa 
## 5          5           3.6          1.4         0.2 setosa
iris.tbl %>%
    filter(Sepal.Length < 7 & Species ==  "setosa") %>% slice(1:5)

iris.tbl %>%
    filter(Sepal.Length < 7,Species ==  "setosa") %>% slice(1:5)

5.10 Summarise observations using dplyr::summarise()

iris.tbl %>% summarise(mean = mean(Petal.Length), sd = sd(Petal.Length))
## # A tibble: 1 x 2
##    mean    sd
##   <dbl> <dbl>
## 1  3.76  1.77
iris.tbl %>% select(-Species) %>%summarise_all(mean)
## # A tibble: 1 x 4
##   Sepal.Length Sepal.Width Petal.Length Petal.Width
##          <dbl>       <dbl>        <dbl>       <dbl>
## 1         5.84        3.06         3.76        1.20
iris.tbl %>% summarise(n = n())
## # A tibble: 1 x 1
##       n
##   <int>
## 1   150

5.11 Summarise observations using dplyr::group_by()

iris.tbl %>% group_by(Species) %>% summarise(mean = mean(Petal.Length), sd = sd(Petal.Length))
## # A tibble: 3 x 3
##   Species     mean    sd
##   <fct>      <dbl> <dbl>
## 1 setosa      1.46 0.174
## 2 versicolor  4.26 0.470
## 3 virginica   5.55 0.552
iris.tbl %>% group_by(Species) %>% summarise(n = n())
## # A tibble: 3 x 2
##   Species        n
##   <fct>      <int>
## 1 setosa        50
## 2 versicolor    50
## 3 virginica     50
iris.tbl %>% group_by(Species) %>% tally()
iris.tbl %>% count(Species)

5.12 Summarise observations using dplyr::group_by()

iris.tbl %>% group_by(Species) %>% filter(Petal.Length >= max(Petal.Length))
## # A tibble: 4 x 5
## # Groups:   Species [3]
##   Sepal.Length Sepal.Width Petal.Length Petal.Width Species   
##          <dbl>       <dbl>        <dbl>       <dbl> <fct>     
## 1          4.8         3.4          1.9         0.2 setosa    
## 2          5.1         3.8          1.9         0.4 setosa    
## 3          6           2.7          5.1         1.6 versicolor
## 4          7.7         2.6          6.9         2.3 virginica
iris.tbl %>% group_by(Species) %>% top_n(1,Petal.Length)
## # A tibble: 4 x 5
## # Groups:   Species [3]
##   Sepal.Length Sepal.Width Petal.Length Petal.Width Species   
##          <dbl>       <dbl>        <dbl>       <dbl> <fct>     
## 1          4.8         3.4          1.9         0.2 setosa    
## 2          5.1         3.8          1.9         0.4 setosa    
## 3          6           2.7          5.1         1.6 versicolor
## 4          7.7         2.6          6.9         2.3 virginica

5.13 Mutating joins using dplyr::*_join()

iris.meta <- tibble(
    Species = factor(c("setosa","versicolor","virginica")),
    Colony = c("A","A","B") ,
    Ploidy = c("diploid","hexaploid","tetraploid"),
    `Common name` = c("Beachhead iris","Harlequin blueflag","Virginia iris")
)
Metadata for iris dataset
Species Colony Ploidy Common name
setosa A diploid Beachhead iris
versicolor A hexaploid Harlequin blueflag
virginica B tetraploid Virginia iris

5.14 Inner join

Inner join: return all rows from x where there are matching values in y, and all columns from x and y. If there are multiple matches between x and y, all combination of the matches are returned.

iris.meta.nosetosa <- iris.meta %>% filter(Species !="setosa")
iris.tbl %>% inner_join(iris.meta.nosetosa,by = "Species") %>% print(n=5)
## # A tibble: 100 x 8
##   Sepal.Length Sepal.Width Petal.Length Petal.Width Species Colony Ploidy
##          <dbl>       <dbl>        <dbl>       <dbl> <fct>   <chr>  <chr> 
## 1          7           3.2          4.7         1.4 versic… A      hexap…
## 2          6.4         3.2          4.5         1.5 versic… A      hexap…
## 3          6.9         3.1          4.9         1.5 versic… A      hexap…
## 4          5.5         2.3          4           1.3 versic… A      hexap…
## 5          6.5         2.8          4.6         1.5 versic… A      hexap…
## # … with 95 more rows, and 1 more variable: `Common name` <chr>

5.15 Outer joins

  • A left join keeps all observations in x.
  • A right join keeps all observations in y.
  • A full join keeps all observations in x and y.

5.16 Left join

Metadata for iris dataset
Species Colony Ploidy Common name
setosa A diploid Beachhead iris
versicolor A hexaploid Harlequin blueflag
virginica B tetraploid Virginia iris

left join:

iris.meta.nosetosa <- iris.meta %>% filter(Species !="setosa")
iris.tbl %>% left_join(iris.meta.nosetosa,by = "Species") %>% print(n=5)
## # A tibble: 150 x 8
##   Sepal.Length Sepal.Width Petal.Length Petal.Width Species Colony Ploidy
##          <dbl>       <dbl>        <dbl>       <dbl> <fct>   <chr>  <chr> 
## 1          5.1         3.5          1.4         0.2 setosa  <NA>   <NA>  
## 2          4.9         3            1.4         0.2 setosa  <NA>   <NA>  
## 3          4.7         3.2          1.3         0.2 setosa  <NA>   <NA>  
## 4          4.6         3.1          1.5         0.2 setosa  <NA>   <NA>  
## 5          5           3.6          1.4         0.2 setosa  <NA>   <NA>  
## # … with 145 more rows, and 1 more variable: `Common name` <chr>

5.17 Right join

iris dataset
Sepal.Length Sepal.Width Petal.Length Petal.Width Species
5.1 3.5 1.4 0.2 setosa
4.9 3.0 1.4 0.2 setosa
7.0 3.2 4.7 1.4 versicolor
6.4 3.2 4.5 1.5 versicolor
6.3 3.3 6.0 2.5 virginica
5.8 2.7 5.1 1.9 virginica

right join:

iris.tbl.nosetosa <- iris.tbl %>% filter(Species !="setosa")
iris.tbl.nosetosa %>% right_join(iris.meta,by = "Species") %>% print(n=5)
## # A tibble: 101 x 8
##   Sepal.Length Sepal.Width Petal.Length Petal.Width Species Colony Ploidy
##          <dbl>       <dbl>        <dbl>       <dbl> <fct>   <chr>  <chr> 
## 1         NA          NA           NA          NA   setosa  A      diplo…
## 2          7           3.2          4.7         1.4 versic… A      hexap…
## 3          6.4         3.2          4.5         1.5 versic… A      hexap…
## 4          6.9         3.1          4.9         1.5 versic… A      hexap…
## 5          5.5         2.3          4           1.3 versic… A      hexap…
## # … with 96 more rows, and 1 more variable: `Common name` <chr>

5.18 Outer joins

full join:

iris.meta.nosetosa <- iris.meta %>% filter(Species !="setosa")
iris.tbl %>% filter(Species != "virginica") %>% full_join(iris.meta.nosetosa,by = "Species")
## # A tibble: 11 x 8
## # Groups:   Species [3]
##    Sepal.Length Sepal.Width Petal.Length Petal.Width Species Colony Ploidy
##           <dbl>       <dbl>        <dbl>       <dbl> <fct>   <chr>  <chr> 
##  1          5.1         3.5          1.4         0.2 setosa  <NA>   <NA>  
##  2          4.9         3            1.4         0.2 setosa  <NA>   <NA>  
##  3          4.7         3.2          1.3         0.2 setosa  <NA>   <NA>  
##  4          4.6         3.1          1.5         0.2 setosa  <NA>   <NA>  
##  5          5           3.6          1.4         0.2 setosa  <NA>   <NA>  
##  6          7           3.2          4.7         1.4 versic… A      hexap…
##  7          6.4         3.2          4.5         1.5 versic… A      hexap…
##  8          6.9         3.1          4.9         1.5 versic… A      hexap…
##  9          5.5         2.3          4           1.3 versic… A      hexap…
## 10          6.5         2.8          4.6         1.5 versic… A      hexap…
## 11         NA          NA           NA          NA   virgin… B      tetra…
## # … with 1 more variable: `Common name` <chr>

5.19 Filtering joins

  • semi_join(x, y) keeps all observations in x that have a match in y.

  • anti_join(x, y) drops all observations in x that have a match in y.

5.20 Filtering joins

Patients files
First_Name Last_Name age sex adress
John Smith 46 M 221B Baker Street
Jane Doe 33 F 57 Rue de Varenne 
treatments.3 %>%
    separate(person,into = c("First_Name","Last_Name"),sep=" ") %>%
    semi_join(treatment.meta,by = c("First_Name","Last_Name"))
##   First_Name Last_Name   treatment result
## 1       John     Smith treatment_a     16
## 2       Jane       Doe treatment_a     23
## 3       John     Smith treatment_b     17
## 4       Jane       Doe treatment_b     25
## 5       John     Smith treatment_c     20
## 6       Jane       Doe treatment_c     23

5.21 Use multiple command together

iris.tbl
## # A tibble: 150 x 5
##    Sepal.Length Sepal.Width Petal.Length Petal.Width Species
##           <dbl>       <dbl>        <dbl>       <dbl> <fct>  
##  1          5.1         3.5          1.4         0.2 setosa 
##  2          4.9         3            1.4         0.2 setosa 
##  3          4.7         3.2          1.3         0.2 setosa 
##  4          4.6         3.1          1.5         0.2 setosa 
##  5          5           3.6          1.4         0.2 setosa 
##  6          5.4         3.9          1.7         0.4 setosa 
##  7          4.6         3.4          1.4         0.3 setosa 
##  8          5           3.4          1.5         0.2 setosa 
##  9          4.4         2.9          1.4         0.2 setosa 
## 10          4.9         3.1          1.5         0.1 setosa 
## # … with 140 more rows

5.22 Use multiple command together

iris.tbl %>%
    gather(key = "Type",value = "obs",-Species) %>%
    separate(Type,into = c("Type_1","Type_2"),se="\\.")
## # A tibble: 600 x 4
##    Species Type_1 Type_2   obs
##    <fct>   <chr>  <chr>  <dbl>
##  1 setosa  Sepal  Length   5.1
##  2 setosa  Sepal  Length   4.9
##  3 setosa  Sepal  Length   4.7
##  4 setosa  Sepal  Length   4.6
##  5 setosa  Sepal  Length   5  
##  6 setosa  Sepal  Length   5.4
##  7 setosa  Sepal  Length   4.6
##  8 setosa  Sepal  Length   5  
##  9 setosa  Sepal  Length   4.4
## 10 setosa  Sepal  Length   4.9
## # … with 590 more rows

5.23 Use multiple command together

iris.tbl %>%
    gather(key = "Type",value = "obs",-Species) %>%
    separate(Type,into = c("Type_1","Type_2"),se="\\.") %>%
    group_by(Species,Type_1,Type_2) %>%
    mutate(med = median(obs))
## # A tibble: 600 x 5
## # Groups:   Species, Type_1, Type_2 [12]
##    Species Type_1 Type_2   obs   med
##    <fct>   <chr>  <chr>  <dbl> <dbl>
##  1 setosa  Sepal  Length   5.1     5
##  2 setosa  Sepal  Length   4.9     5
##  3 setosa  Sepal  Length   4.7     5
##  4 setosa  Sepal  Length   4.6     5
##  5 setosa  Sepal  Length   5       5
##  6 setosa  Sepal  Length   5.4     5
##  7 setosa  Sepal  Length   4.6     5
##  8 setosa  Sepal  Length   5       5
##  9 setosa  Sepal  Length   4.4     5
## 10 setosa  Sepal  Length   4.9     5
## # … with 590 more rows

5.24 Use multiple command together

iris.tbl %>%
    gather(key = "Type",value = "obs",-Species) %>%
    separate(Type,into = c("Type_1","Type_2"),se="\\.") %>%
    group_by(Species,Type_1,Type_2) %>%
    mutate(med = median(obs)) %>%
    mutate(Size = ifelse(obs < med,"Small","Big")) %>%
    left_join(iris.meta,by = "Species")
## # A tibble: 600 x 9
## # Groups:   Species, Type_1, Type_2 [?]
##    Species Type_1 Type_2   obs   med Size  Colony Ploidy  `Common name` 
##    <fct>   <chr>  <chr>  <dbl> <dbl> <chr> <chr>  <chr>   <chr>         
##  1 setosa  Sepal  Length   5.1     5 Big   A      diploid Beachhead iris
##  2 setosa  Sepal  Length   4.9     5 Small A      diploid Beachhead iris
##  3 setosa  Sepal  Length   4.7     5 Small A      diploid Beachhead iris
##  4 setosa  Sepal  Length   4.6     5 Small A      diploid Beachhead iris
##  5 setosa  Sepal  Length   5       5 Big   A      diploid Beachhead iris
##  6 setosa  Sepal  Length   5.4     5 Big   A      diploid Beachhead iris
##  7 setosa  Sepal  Length   4.6     5 Small A      diploid Beachhead iris
##  8 setosa  Sepal  Length   5       5 Big   A      diploid Beachhead iris
##  9 setosa  Sepal  Length   4.4     5 Small A      diploid Beachhead iris
## 10 setosa  Sepal  Length   4.9     5 Small A      diploid Beachhead iris
## # … with 590 more rows

5.25 Use multiple command together

iris.tbl %>%
    gather(key = "Type",value = "obs",-Species) %>%
    separate(Type,into = c("Type_1","Type_2"),se="\\.") %>%
    group_by(Species,Type_1,Type_2) %>%
    mutate(med = median(obs)) %>%
    mutate(Size = ifelse(obs < med,"Small","Big")) %>%
    left_join(iris.meta,by = "Species") %>%
    arrange(desc(obs)) %>% slice(1)
## # A tibble: 12 x 9
## # Groups:   Species, Type_1, Type_2 [12]
##    Species   Type_1 Type_2   obs   med Size  Colony Ploidy  `Common name`  
##    <fct>     <chr>  <chr>  <dbl> <dbl> <chr> <chr>  <chr>   <chr>          
##  1 setosa    Petal  Length   1.9  1.5  Big   A      diploid Beachhead iris 
##  2 setosa    Petal  Width    0.6  0.2  Big   A      diploid Beachhead iris 
##  3 setosa    Sepal  Length   5.8  5    Big   A      diploid Beachhead iris 
##  4 setosa    Sepal  Width    4.4  3.4  Big   A      diploid Beachhead iris 
##  5 versicol… Petal  Length   5.1  4.35 Big   A      hexapl… Harlequin blue…
##  6 versicol… Petal  Width    1.8  1.3  Big   A      hexapl… Harlequin blue…
##  7 versicol… Sepal  Length   7    5.9  Big   A      hexapl… Harlequin blue…
##  8 versicol… Sepal  Width    3.4  2.8  Big   A      hexapl… Harlequin blue…
##  9 virginica Petal  Length   6.9  5.55 Big   B      tetrap… Virginia iris  
## 10 virginica Petal  Width    2.5  2    Big   B      tetrap… Virginia iris  
## 11 virginica Sepal  Length   7.9  6.5  Big   B      tetrap… Virginia iris  
## 12 virginica Sepal  Width    3.8  3    Big   B      tetrap… Virginia iris

6 Data visualisation using ggplot2

6.1 A base plot

plot(Petal.Width~Petal.Length,data=iris.tbl)

6.2 A ggplot2 plot

iris.tbl %>% ggplot(aes(x=Petal.Width,y=Petal.Length)) +
    geom_point()

6.3 A ggplot2 plot

iris.tbl %>% ggplot(aes(x=Petal.Width,y=Petal.Length)) +
    geom_point() +
    geom_smooth(method = "lm")

6.4 The three key components of every ggplot2

The three components of every graphics :

  • data: the dataset who’s need to be plotted.
  • aesthetics: aesthetics mapping between variables in the data (x,y,visuals properties, …).
  • geoms: one or more layers to render each observations.

6.5 The three key components of every ggplot2

The three components of every graphics :

  • data: the dataset who’s need to be plotted.
  • aesthetics: aesthetics mapping between variables in the data (x,y,visuals properties, …).
  • geoms: one or more layers to render each observations.
p <- iris.tbl %>%# Data
    ggplot(aes(x=Petal.Width,y=Petal.Length)) +# aesthetics
    geom_point()# Layer: points

Add layer : smooth with linear model

p + geom_smooth(method = "lm")

More you add components with +, more you’ll build sophisticated plots.

6.6 Add more aesthetics

p <- iris.tbl %>%# Data
    ggplot(aes(x=Petal.Width,y=Petal.Length,col=Species)) +# aesthetics
    geom_point()# Layer: points
print(p)

6.7 Add more aesthetics

p + geom_smooth(method = "lm")

6.8 Add more aesthetics

p <- iris.tbl %>%# Data
    ggplot(aes(x=Petal.Width,y=Petal.Length)) +# aesthetics
    geom_point(aes(col=Species)) +# Layer: points
    geom_smooth(method = "lm")
print(p)

6.9 Other aesthetics attributes

p <- iris.tbl %>%# Data
    ggplot(aes(x=Petal.Width,y=Petal.Length)) +# aesthetics
    geom_point(aes(shape=Species))# Layer: points
print(p)

6.10 Other aesthetics attributes

p <- iris.tbl %>%# Data
    ggplot(aes(x=Petal.Width,y=Petal.Length)) +# aesthetics
    geom_point(aes(col=Species,size = Sepal.Length))# Layer: points
print(p)

6.11 Use fixed value

p <- iris.tbl %>%# Data
    ggplot(aes(x=Petal.Width,y=Petal.Length)) +# aesthetics
    geom_point(col="blue",size=4) +# Layer: points
    geom_smooth(method = "lm",col="red",fill="orange")
print(p)

6.12 Plot geoms

6.13 Facetting

Facetting:

  • facet_grid(variable~foliaces) : Display all possibility even if some plots are empty.
  • facet_wrap(variable~foliaces) : Display only the plots having actual values.

6.14 Graphical options

p <- iris.tbl %>%# Data
    ggplot(aes(x=Petal.Width,y=Petal.Length)) +# aesthetics
    geom_point(aes(col=Species,size = Sepal.Length)) +# Layer: points
    scale_color_brewer(palette="Set3") #A new palette
print(p)

6.15 Graphical options

p <- iris.tbl %>%# Data
    ggplot(aes(x=Petal.Width,y=Petal.Length)) +# aesthetics
    geom_point(aes(col=Species,size = Sepal.Length)) +# Layer: points
    scale_color_manual(values = c("red","blue","green"),  #A new palette
                       labels = c("Setosa","Versicolor","Virginica"))
print(p)

6.16 Themes

6.17 Themes

6.18 Themes

6.19 Extensions

  • ggplot2 extensions: This site tracks and lists ggplot2 extensions developed by R users in the community. https://www.ggplot2-exts.org/.
  • cowplot: Arranging graphs into a grid and improve plot design.
  • ggpubr & __ ggstatsplot__: Publication ready plots & add statistical tests.
  • ggrepel: Automatic label placement.
  • esquisse: Explore and Visualize Your Data Interactively with ggplot2.
  • ggvis: Create rich interactive graphics with a syntax similar in spirit to ggplot2.

6.20 Why I use ggplot2

  • The “default” output is much nicer than with base graphics.
  • Automatic legend, color with mapping to the variables.
  • Easy to combine multiple plot.
  • Easy facetting.
  • A lot of themes.
  • A lot of extensions …